Automated Asset Management - September 16, 2016

The digitalisation of the asset management industry will mostly impact distribution models

In this interview, we talk to Bernd Scherer, Head of Quantitative Strategies at Deutsche Asset Management and Research Associate with the EDHEC-Risk Institute. We discuss his new position paper, “What Investment Robots Need To Know”, as well as the impacts and implications automated asset management offerings have on the industry, and we take a deeper look into his research projects.

Bernd Scherer

You joined the EDHEC-Risk Institute as a research associate at the beginning of this year. Could you tell us why you accepted to join EDHEC's research team?

Bernd Scherer: industry practitioners who are committed to all aspects of the asset management business. The group’s combined expertise and research output across regulatory, risk management or product-related research offers me a significant advantage in both my academic work and my understanding of an ever-changing industry. And sometimes I can even make my own modest contribution. What attracts me most to EDHEC is their relentless focus on themes that are highly relevant to asset owners as well as to asset managers. And most of the time EDHEC is at the forefront of identifying topics that will prove to be important in the near future. The reason EDHEC is able to do this is precisely down to this network of researchers that provides an excellent feedback mechanism. In my view, EDHEC-Risk is performing the most important task of an academic institution: providing knowledge almost as a public good to the industry. In this role, I also see EDHEC as a powerful rational voice in a world blurred by political interest (Solvency 2, shorting bans, etc.).

You have just written an EDHEC-Risk Paper entitled “What Investment Robots Need To Know – Evidence Survey Data from Panel”. Could you tell us about the empirical model you use? Give us your observations and recommendations?

Bernd Scherer: There are three main challenges. The first one concerns the definition of an portfolios to individual investors based on investor characteristics such as age, net income or self-assessments of risk aversion. Using a new German household panel data, we investigate the key household characteristics that drive private asset allocation decisions. This information allows us to assess which set of variables should be included in algorithmic portfolio advice. Using heavily cross-validated classification trees, we find that a combination of household balance sheet variables – describing the ability to take risks (e.g. net wealth) – and household personal characteristics – describing the willingness to take risks (e.g. risk aversion) – best explain the cross-sectional variation in household portfolio choice. Our empirical evidence is in line with published work on normative portfolio choice including outside wealth or shadow assets/liabilities. The presented results suggest a more holistic modelling of household characteristics. Including background risks in the form of household leverage does make obvious investment sense. Unfortunately, few robo-advisors truly model economic household balance sheets. While they often ask questions somewhat related to household net wealth or shadow liabilities, this information does not enter their models, i.e. it does not change recommended allocations. Unless this changes, most offerings do not offer truly individualised advice. In my view, this should change in the interest of clients.

The wealth management industry is about to experience a new industrial revolution where robo-advice and fintech will play an important role. How do you see the impacts & implications of these automated asset management offerings on the industry?

Bernd Scherer: In my view, the digitalisation of the asset management industry will mostly impact distribution models, particularly when low interest rates make typical asset management fees look oversized. I am, however, not so sure standalone robo-advisors will be a long term commercial success. First, I believe the costs per account are too high for the bulk of the many small accounts (€10,000) robo-advice is targeted for. A fee of 0.5% on €10,000 amounts to a mere €50 per year. This hardly covers the costs of running an automated execution and fund management platform. Second, the fee model itself looks unsustainable. Why should anyone pay a 0.5% fee for advice on how to structure a €500,000 investment? Instead, one would invest €5,000 and replicate the advice (usually buying a portfolio of ETFs) on the remaining €495,000 for free. This would save around €2,500 every year. Instead, I believe asset managers and banks will need to digitalise (automate) all aspects of their balance sheets to leverage their business. Unless robo-advisors also move to production, I don’t anticipate them being a threat to traditional suppliers of asset management services. This might change if technology allows them to also disrupt production. Imagine investors being able to disintermediate ETFs by investing in low cost individualised baskets via trading platforms.

What is your current research focus, and what are the main research projects that you are lining up for 2016/2017?

Bernd Scherer: As Head of Quantitative Strategies for Deutsche Asset Management, right now I am mostly interested in multi-factor investing, i.e. creating portfolios of risk premia, diversified across asset classes and investment themes. This creates many interesting academic research questions relevant for creating an attractive product. How do you define what constitutes the risk premium investment universe? How can we allocate among risk premia taking into account the stylised effects of their return distributions (skew, volatility and returns clustering, tail correlation, etc.). How can we explain/predict the return variation (cyclicality) in risk premia? How do risk factor allocations change for different asset owners? What are the risk factors embedded in client liabilities? Apart from this, I continue to be interested in the world of robo-advice. In particular, my research centres on building an extensive model of household balance sheets, guided by panel survey data and how to use this information in normative models of portfolio choice (what investors should do), i.e. how can we build a truly individualised robo-advisor?

About Bernd Scherer

EDHEC-Risk and visiting professor at WU Vienna with 23 years of industry experience. Before re-joining Deutsche in 2015, Bernd was MD at Morgan Stanley (London), Professor of Finance at EDHEC Business School (Nice), Head of Research at DeAM (New York), Chief Scientific Officer at First Private (Frankfurt) and Chief Investment Officer at FTC Capital (Vienna). He combines extensive investment experience across all asset classes with strong academic credentials. His academic work has been published in journals like the Journal of Banking and Finance, Journal of Financial Markets, Journal of Empirical Finance, Journal of Economics and Statistics, Quantitative Finance, Journal of Derivatives, Journal of Portfolio Management, Financial Analysts Journal, Journal of Investment Management, Risk, Financial Markets and Portfolio Management, Journal of Asset Management, etc. Bernd is also author/editor of eight books on quantitative asset management for Risk, Springer and Oxford University Press and occasionally writes for the Financial Times.

URL for this document:

Hyperlinks in this document: